Low-Greenhouse-Gas Consumption Strategies and Impacts on ...

Stockholm Environment Institute, Working Paper 2012-01

Low-Greenhouse-Gas Consumption Strategies and Impacts on Developing Countries Peter Erickson, Anne Owen and Ellie Dawkins

Stockholm Environment Institute Kräftriket 2B 106 91 Stockholm Sweden Tel: +46 8 674 7070 Fax: +46 8 674 7020 Web: www.sei-international.org Author contact: Peter Erickson Stockholm Environment Institute-U.S. Centre 1402 Third Avenue, Suite 900 Seattle, WA 98101, USA [email protected] Head of Communications: Robert Watt Publications Manager: Erik Willis Cover Photo: Workers in a clothes factory in Dharavi slum in Mumbai, India, sew blouses for export © Flickr/BBC World Service. This publication may be reproduced in whole or in part and in any form for educational or non-profit purposes, without special permission from the copyright holder(s) provided acknowledgement of the source is made. No use of this publication may be made for resale or other commercial purpose, without the written permission of the copyright holder(s). About SEI Working Papers: The SEI working paper series aims to expand and accelerate the availability of our research, stimulate discussion, and elicit feedback. SEI working papers are work in progress and typically contain preliminary research, analysis, findings, and recommendations. Many SEI working papers are drafts that will be subsequently revised for a refereed journal or book. Other papers share timely and innovative knowledge that we consider valuable and policyrelevant, but which may not be intended for later publication. Copyright © March 2012 by Stockholm Environment Institute

STOCKHOLM ENVIRONMENT INSTITUTE WORKING PAPER NO. 2012-01

Low-Greenhouse-Gas Consumption Strategies and Impacts on Developing Countries

Peter Erickson, Anne Owen 1 and Ellie Dawkins Stockholm Environment Institute

ABSTRACT A growing body of research shows how shifts in consumer behaviour could lead to reductions in greenhouse gas (GHG) emissions. By buying less “stuff” and fewer high-GHG items (e.g. red meat), and redirecting any spending to low-GHG alternatives, consumers could help reduce emissions. Altogether, these shifts could reduce emissions associated with consumption in highincome countries by at least 10 per cent, and likely more. Many of the goods consumed in high-income countries are produced in low-income countries, however, raising questions about the economic impact of reduced consumption on those countries. Starting with the United Kingdom as an example, we analyse the potential economic impacts of low-GHG consumption strategies. We find that if the U.K. and all other high-income countries shifted spending to lower GHG products and services, lower-income countries would be disproportionately affected, with average GDP losses greater than 5 per cent in the world’s poorest countries. These findings raise a flag of caution about how to pursue low-GHG consumption in highincome countries. International trade can raise incomes and standards of living in developing countries. Measures that reduce trade –under the banner of low-GHG or “sustainable” consumption or in other ways, such as by promoting local purchasing – can hurt poorer countries that critically depend on that income. It is possible, however, that by preferentially sourcing products from low-GHG and low-income regions, high-income countries could foster both GHG and development benefits. Further work is needed to identify specific opportunities, taking into account factors – such as marginal energy sources and production practices – that affect the GHG-intensity of increased production in low-income countries.

1

Current affiliation: University of Leeds

LOW-GREENHOUSE-GAS CONSUMPTION STRATEGIES AND IMPACTS ON DEVELOPING COUNTRIES

SEI WP 2012-01

CONTENTS 1. Introduction and Context ............................................................................................ 3 2. Methodology ............................................................................................................... 4 Defining a Set of Low-GHG Behaviours ............................................................................. 4 Modeling Impacts of Behaviours with an MRIO Model ....................................................... 7 3. Results ................................................................................................................. 10 Limitations of Analysis ..................................................................................................... 13 4. Discussion: Options for Addressing Development in Sustainable Consumption ...... 14 Purchase from Poor and Low-GHG Countries .................................................................. 15 Import Higher-Cost, Higher-Quality, and Value-Added Goods......................................... 16 Lower the GHG-Intensity of Production in Poor Countries ................................................. 17 Summary ........................................................................................................................ 17 5. Product Focus: Clothing ........................................................................................ 18 Comparing GHG-Intensity across Countries .................................................................... 19 Priorities for Further Research on Shifting Location of Production ...................................... 22 6. Conclusions .............................................................................................................. 23 Acknowledgments ........................................................................................................ 24 References .................................................................................................................... 25

2


SEI WP 2012-01

1. INTRODUCTION AND CONTEXT Greenhouse gases are accumulating in the atmosphere to levels that could dangerously disrupt the global climate system (IPCC 2007). As nations have debated how far to reduce greenhouse gas emissions, what mechanisms to employ, and how to distribute responsibility, analysts have created scenarios of how particular technologies and measures could reduce emissions over time. These scenarios, developed by international quasi-governmental organisations (e.g., the Intergovernmental Panel on Climate Change or the International Energy Agency), research institutions, and private-sector analysts, have played important roles in demonstrating what levels of GHG emission reduction are possible. Most emission reduction scenarios have focused especially on widespread changes in energy infrastructure and technologies, such as how electricity is produced and what technologies or fuels are used to make goods, move people and goods, or heat and cool buildings (IEA 2010; WWF 2011; McKinsey & Company 2010; Metz et al. 2007). In constructing these scenarios, analysts largely focus on how to provide goods, services, and other amenities with fewer emissions, rather than on changing the mix of goods and services consumed. In other words, most mitigation scenarios focus largely on shifts in production, rather than in consumption (except for behavioural responses to a price on carbon, typically in the form of reducing the direct use of fossil fuels – e.g. less driving). In general, the same types of goods and services are consumed, at roughly the same levels, in mitigation scenarios as in the corresponding business-as-usual scenarios. However, limiting warming to 2°C or 1.5°C, a widely embraced policy goal (United Nations Framework Convention on Climate Change 2011a), and averting the worst impacts of climate change may require broader shifts in what and how we consume. While widespread adoption of low-GHG technology will be essential, shifts to low-GHG lifestyles may also be needed. This realisation, coupled with lagging progress in international negotiations and within key countries (e.g., the United States), has helped bring a resurgence of interest in consumptionbased approaches for climate mitigation. For example, several recent bottom-up studies have quantified the potential impacts of consumption-based actions on greenhouse gas emissions. These studies demonstrate that low-GHG consumption behaviours, such as reduced consumption of red meat, can contribute substantially to global GHG emissions abatement (Scott et al. 2009; Dietz et al. 2009; Jones and Kammen 2011; BioRegional and London Sustainable Development Commission 2009; Stehfest et al. 2009). The goal of this paper is to reflect on some possible global economic implications of lowGHG consumption. In particular, we explore how shifting to low-GHG consumption, as defined by recent assessments of its role in climate mitigation, may or may not contribute to another key objective of sustainable development: global development and poverty alleviation. For guidance, we take a cue from the United Nations Earth Summit, held in Rio de Janeiro in 1992, which brought sustainable consumption and production to the world stage, and which cited poverty alleviation as a key “basis for action”. From the UN’s Agenda 21: …the major cause of the continued deterioration of the global environment is the unsustainable pattern of consumption and production, particularly in industrialised countries, which is a matter of grave concern, aggravating poverty and imbalances… (United Nations 1992). According to Agenda 21, over-consumption in industrialised countries is depriving poor nations of access to resources. At the same time, poor countries benefit from consumption in 3


SEI WP 2012-01

industrialised countries through trade of raw materials and products. As international trade has grown substantially in the nearly 20 years since the Earth Summit, 2 the links between consumption in high-income countries and development in lower-income exporter countries are even more relevant. Low-income countries have become more reliant on trade – with exports comprising 12 per cent of their GDP in 1992, 20 per cent in 2010, and as much as 22 per cent immediately before the recent global recession (World Bank 2011). These trends helped lead to higher standards of living in increasingly export-oriented countries (Irwin and Terviö 2002), and create a dilemma: If high-consumption countries were to decrease trade in the course of pursuing low-GHG consumption, 3 what would be the effect on their low-income trading partners? This is the question we explore in this paper. Other researchers have presented a similar case for individual products, such as strawberries from Africa (Müller 2007); here, we instead look at a broad range of measures, and bring a quantitative perspective that we have not yet seen in the literature. We first describe our methodology, which includes a review of scenarios of low-GHG consumption in higher-income countries, selection of one particular scenario for further analysis, and use of a multi-regional input-output (MRIO) model of the global economy. We then present results of our analysis, showing the impacts of low-GHG consumption in highincome countries on per-capita incomes in low-income countries. Lastly, we discuss potential options for research and practice on low-GHG consumption patterns that can simultaneously reduce global GHG emissions and increase development benefits in developing countries. We should note that while we raise the question of consumption here, we do so only in terms of type, not overall levels. In other words, we keep overall consumption levels constant in the United Kingdom and other high-income countries (redirecting any freed-up income in our analysis to other, low-GHG services) and do not attempt to address pathways or global economic effects of reduced overall consumption (such as Jackson 2009) in this analysis. 2. METHODOLOGY To explore trade impacts on lower-income countries, we first define a set of low-GHG consumption behaviours that have gained traction in recent analyses. We then explore the trade-related impacts of these behaviours on lower-income countries, defined here as all countries that the World Bank categorises as low or medium-income. 4 Our primary tool for analysis is a global, multi-regional input-output (MRIO) model, a type of tool that has become common in assessing the GHG and economic impacts of trade (Minx et al. 2009; Wiedmann et al. 2011). These steps are described further below. Defining a Set of Low-GHG Behaviours

The focus of this analysis is on consumer purchasing behaviours, which have drawn increasing attention in recent years as a means to reduce GHGs. Consumer purchasing also presents an opportunity for development impacts, due to the substantial international trade of consumer goods produced in lower-income countries.

2

According to the World Bank (2011), trade (exports) represented 15 per cent of global GDP in 1972, 20 per cent in 1992, and 28 per cent in 2010. 3 High-consumption countries could also reduce or change consumption in response to other factors or trends, such as buying “local” (a theme we will return to at the end of the paper) or due to other environmental burdens beyond GHGs, such as water or toxics. We do not evaluate other motivations for changing consumption, or other environmental burdens or limits, in this paper. 4 See http://data.worldbank.org/about/country-classifications.

4


SEI WP 2012-01

Our focus is on shifts in consumption that extend beyond the direct use of energy, such as diet shift, reduction in food waste, longer product lifespans, and overall reduced purchasing of “stuff”. The reason for this focus is that although there is widespread agreement that changes in energy production and consumption will be necessary to address climate change, the role of other consumer choices is less clear, and the potential economic impacts on lower-income countries are, arguably, larger. (Lower-income countries produce many goods and services for high-income countries, but those other than fossil fuels yield about six times more economic added value than fossil fuels. 5) We know of no widespread agreement (implicit or explicit) on what behaviour changes could have significant impacts on global GHGs. Whereas measures that address energy supply and efficiency (among others) have been well-chronicled by the Intergovernmental Panel on Climate Change and others and form the building blocks of most bottom-up scenarios of GHG abatement, we know of no similar typology of consumer behaviour changes on goods and services. Analyses of diets and food are perhaps most common (Stehfest et al. 2009; Audsley et al. 2009; Garnett 2011; Brohmann and Barth 2011), but other assessments have looked at different suites of measures and behaviour shifts, such as reduction in purchases of clothing (Allwood et al. 2006; Carbon Trust 2011). To explore what categories of behaviour changes are commonly discussed in the literature, we reviewed several recent scenarios that analysed behavioural measures and the resulting magnitudes of GHG reductions. Since our intent is to assess the impacts of an entire group of measures, we focused our review on studies that addressed more than one product category (Table 1).

5

Source: Author analysis of the data in the underlying MRIO model.

5


SEI WP 2012-01

Table 1. Comparison of Bottom-up Consumption-based Scenarios Scenario:

Categories Addressed Food

EU – European Commission (Brohmann and Barth 2011)

X

U.K. – London (BioRegional and London Sustainable Development Commission 2009)

U.K. – University of Surrey (Druckman and Jackson 2010)

U.K. – WRAP (Scott et al. 2009)

US – Michigan State and others (Dietz et al. 2009)

US – UC Berkeley (Jones and Kammen 2011)

X

X

X

Goods

X

X

X

Services Construction

X X

X X

X X

X

X

X

X

X

X

X

X

Personal Transportation Home Energy

X

GHG Reductions Due to Behaviour Shifts

6

~14% 8

~16%

1014% 9

N/A7

4% (food only)

Personal Transport, Home Energy

~20%

~21%

N/A

N/A

16%

Total

~34%

~37%

10-14%

N/A

20%

London

U.K.

U.K.

US

US

HH, Gov

HH

HH, Gov Through 2050

HH

HH

Through 2015

None

Yes

No

No

Food, Goods, Services, Construction

N/A 7

X

Other Factors Geographic Focus

EU-27

Institution Focus Time scale

Through 2030

Through 2050

None

Quantified rebound

No

No

No

Based on this review, it is clear that analysts have not settled on a standard set of behaviour shifts. This complicates efforts to compare the assumptions and results of studies – some of which focus on nearly all forms of consumption, whereas others focus on a particular subset. In Table 1, we summarise the level of GHG-reductions that studies suggest is achievable via behavioural measures. Studies that looked at measures addressing food, goods, services, and construction yielded GHG reductions in the range of 14 to 16 per cent (before considering rebound) combined from these measures. Measures that address personal transport and home energy yielded GHG reductions of 16 to 21 per cent from these measures. Of the scenarios reviewed in Table 1, the WRAP scenario for the U.K. (Scott et al. 2009) is the only scenario reviewed that both extends across the non-energy categories of consumer purchasing (e.g., food, goods, and services) and also includes consideration of how income freed up by reduced consumption might be redirected (i.e. the rebound effect). Including the rebound effect is important, because shifts in consumer behaviour (e.g. reduced purchasing of 6

At maximum implementation of measures considered, regardless of what year that implementation occurs. Source did not report a total BAU against which to assess the impact of the measures assessed. 8 Approximated based on identification of consumption-based measures in the study’s appendix. Reductions would likely be greater if the study did not also apply aggressive efficiency improvements in most sectors. Range reflects measures that only apply to goods (including food) and services (on the low end) to also including measures that address personal transportation and home energy (on the high end). 9 Depending on the degree of rebound. 7

6


SEI WP 2012-01

a high-GHG good) would likely enable other purchases (Hertwich 2008; Girod et al. 2010; Ornetzeder et al. 2008). Because of this rebound effect, the GHG benefits could be muted, and economic impacts could also differ. 10 We note that the WRAP scenario also finds GHG reductions from shifting behaviours associated with food, goods, services, and construction that are in line with the other studies reviewed, and so can be used to represent the scale of potential shifts to low-GHG consumption. For these reasons, we select the WRAP scenario as the set of behaviour changes to model for this analysis. The WRAP scenario, released by the Waste & Resources Action Programme in 2009, charted the long-term (through 2050) GHG abatement potential behaviour change in the U.K., and was the most detailed accounting of the embodied GHGs in U.K. consumption at the time (Scott et al. 2009). 11 Among the specific behaviours considered were reduction in food waste, shifts in diet, using goods (such as clothing and home furnishings) longer, and shifting from goods to services for select goods (e.g. sharing services and rentals of high-end clothing, glass and tableware, household tools, personal vehicles). The strategies considered in the most aggressive variant of the WRAP scenario, used here, are listed in Table 2. The scenario found (after freed-up spending was redirected to low-GHG services, via the “rebound effect”) that aggressive implementation of these strategies could reduce the emissions associated with U.K. consumption by 10 per cent, compared to baseline emissions in 2050. Table 2. Behaviour Changes included in WRAP Scenario ‘Beyond Best Practice’ Variant Category

Measure

Food

All edible food waste is eliminated, reducing the need for food purchases Meat and dairy consumption declines 75% (with no replacement with other food) Goods that are still working are no longer discarded:; 90% of goods used to their full “technological lifespan” Durability of goods increases, reducing need for new goods by 40% Several goods are shared (rented) rather than owned personally: clothing, glassware and tableware, tools and equipment, vehicles, and recreational and audio-visual equipment (with rental rates varying by good) The GHG intensity of government procurement declines 90% Shift from goods to services (i.e., shared goods), per above 90% of homes slated for demolition are brought back into use, reducing the need for new builds 12

Goods

Services Construction

Modelling Impacts of Behaviours with an MRIO Model

To model the trade-related impacts of the list of behaviours identified, we used environmentally extended input-output (IO) techniques similar to those used in the original WRAP analysis. Instead of working with the exact model used in the WRAP study (which had a limited ability to distinguish world regions), we use an updated multi-regional Input-

10

The WRAP scenario does not, however, address another kind of rebound effect – that due to price effects. For example, if high-consumption countries were to reduce consumption of a particular good, the global price of this good could be expected to decline, enabling greater consumption of it by low-income countries who could then better afford it. We do not explicitly consider this possibility here, because our primary focus is not the GHG impacts of shifts to low-GHG consumption, but rather the resulting trade flows from low-income to highconsumption countries, which would not be affected by this indirect rebound effect. 11 The study has since been published in the peer-reviewed literature as Barrett and Scott (2012). 12 In the WRAP study, this was considered a supply- (production-) side measure, not demand- (behaviour-) side measure, but we list it here as a behaviour change since it is reducing the demand for new housing. However, following the WRAP definition, we do not consider it in the remainder of the analysis.

7


SEI WP 2012-01

Output (MRIO) model that allows us to better identify the countries (and sectors) in which trade impacts could occur. 13 IO models have become standard tools for assessing the distribution of an industry’s product throughout the economy, including by country and economic sector (Miller and Blair 2009; Wiedmann et al. 2011). In recent years, these tools have been extended to cover global economies and focus on GHG impacts, leading to a number of assessments of the GHGs embodied in trade (Peters and Hertwich 2008; Hertwich and Peters 2009; Peters et al. 2011; Wiedmann 2008). The models are subject to several uncertainties, including those that arise from aggregating hundreds of individual sectors and regions into smaller, more manageable numbers; harmonising trade data across regions; and converting currencies into a common unit (Lenzen et al. 2010). These and other sources of uncertainty and sensitivity are discussed extensively in the literature (Lenzen et al. 2010; Weber 2008) and so are not described in detail here. However, in general, uncertainties are lowest at the level of entire economies (e.g., all emissions embodied in the U.K.’s consumption) and highest for particular trade flows between pairs of countries. To model the shifts in behaviour, we obtained the raw data from the WRAP study authors (Scott et al. 2009) in terms of expenditures by category in both the baseline and “beyond best practice” (mitigation) scenarios. 14 Table 3 displays the change in spending in each category. Note that reductions in spending in some sectors (e.g., bovine meat products, dairy products, clothing) are profound, and redirecting these savings leads to a large increase in spending on low-GHG recreational services. We then apply these reductions in spending to our MRIO model for spending in our model’s analysis year (2004), to estimate the scale of trade flows and GHGs targeted in the WRAP scenario. Although this approach obscures the many years it would take to achieve such societal shifts, as well as other trends that would affect emissions over time (such as changes in production technologies), 15 this approach provides a simple estimate of the scale of trade flows (and emissions) targeted. Table 3 summarises actual (2004) mitigation (WRAP) scenario expenditures for the sectors affected, along with the fraction of final demand of each commodity that is satisfied by foreign production. Total final demand in the U.K. is assumed to remain unchanged, as all reductions in spending are assumed to be “rebounded” to recreational and other services.

13

The model is based on the GTAP 7 database and therefore has 113 world regions and a base year of 2004. It is based on one created at the Norwegian University of Science and Technology (Hertwich and Peters 2010), as used in the European Union’s EUREAPA online tool (http://www.oneplaneteconomynetwork.org/eureapa.html). 14 For the purposes of the WRAP study, the authors used 123 disaggregated sectors, defined using the Standard Industrial Classification (SIC) system (Scott et al. 2009). Since the MRIO model we used is based on underlying data of the Global Trade and Analysis Project (GTAP, version 7), we then constructed a correspondence table between SIC sector codes to GTAP codes. 15 This method considers only the impact of shifting consumption relative to the baseline from the WRAP scenario, in order to explore the impact of these measures in isolation; any other background changes within the scenario, such as economic or population growth, are excluded.

8


SEI WP 2012-01

Table 3. Final Demand, Trade Volumes, and WRAP Scenario Shifts

Consuming Sector ID (GTAP)

Consuming Sector Name (GTAP)

Final Demand 16 (2004), Million USD

1

Paddy rice

4

Vegetables fruit nuts

8

Crops nec

4,050

10

Animal products nec

1,156

12

Wool silk-worm cocoons

19

Bovine meat products

4,088

20

Meat products nec

6,224

21

Vegetable oils and fats

1,431

22

Dairy products

14,505

23

Processed rice

289

24

Sugar

25

Food products nec

82,432

27

Textiles

26,433

28

Wearing apparel

50,711

29

Leather products

9,870

30

Wood products

3,575

31

Paper products publishing

32,048

33

Chemical rubber plastic products

48,924

34

Mineral products nec

37

Metal products

15,297

38

Motor vehicles and parts

92,783

39

Transport equipment nec

12,285

40

Electronic equipment

51,384

41

Machinery and equipment nec

66,648

42

Manufactures nec

57,629

55

Recreational and other services

77,985

All Others

62 12,208

47

1,687

6,406

All Others

1,296,104

Total

1,930,515

Fraction of Final Demand Satisfied by Foreign Production (2004) 100% 57% 46% 23% 98% 35% 53% 49% 24% 51% 38% 23% 42% 35% 55% 65% 24% 47% 32% 34% 59% 40% 62% 54% 34% 17% 17% 24%

WRAP Scenario, Reduction Compared to Baseline (including Rebound) -15% -1% 0% -21% -9% -87% -87% -17% -87% -16% -17% -33% -84% -87% -74% -74% -32% -1% -78% -78% -92% -25% -84% -30% -65% +440% 0% 0%

To implement the consumption shift in the MRIO model, we assume that any change in consumption occurs equally for domestic and foreign components of demand. In the case of the example above, if red meat expenditures are reduced by 87 per cent, we assume that demand for both domestic and imported red meat drops by 87 per cent. Within the MRIO, we track how a change in spending on a product affects every contributing producing sector from every region throughout the product’s supply chain. In other words, an 87 per cent reduction in red meat consumption will not only reduce demand for cattle but also, in turn, reduce demand for energy, fertilisers, and pesticides needed to grow feed for cattle, regardless of where in the world those items are produced. An 87 per cent reduction in spending on red 16

Final demand is spending by households, government, business capital investment, and net changes to stocks.

9


SEI WP 2012-01

meat therefore translates to an 87 per cent reduction in GHGs associated with red meat. We use the MRIO model to follow this through the entire supply chain, from country to country and sector to sector. Using this MRIO structure, we can estimate changes in each country’s GDP (and greenhouse gas emissions) based on changes in consumption in the U.K. or in any other group of countries. For further details on construction of the underlying MRIO model, please see Hertwich and Peters (2010). We analysed the impacts of shifting consumption in the U.K. on four country groupings: 

The U.K. itself, to look at domestic impacts of the behaviour changes;



High-income (outside the U.K.);



Lower-income;



Least Developed Countries (LDCs), or world’s poorest countries, which are a subset of the lower-income countries.17

Table 4 provides some descriptive statistics of the regions analysed. Note that 76 per cent of the U.K.’s consumption is satisfied by domestic production. Table 4. Descriptive Statistics of Regions Analysed (2004) Region

U.K. High-income (ex U.K.) Lower-income I LDCs17 Total (World)

GDP 18 (2004 US $)

GDP/Capita (2004 US $)

Value Added for U.K. Consumption (2004 US $)

Value Added as Fraction of U.K. Consumption

60,000

2,124,000

35,400

1,461,000

76%

1,192,000 5,371,000

30,598,000 8,248,000

31,642 1,536

368,000 102,000

19% 6%

Population

374,000

114,000

310

2,000